Hi, I’m Mikhail. I’m exploring how AI can help with real work, real life, and the messy problems in between.

This site is where I share what I’m learning, building, and testing: small AI-assisted systems, practical automations, and personal tools that solve real problems for real people.

I’m not an AI guru or a professional software engineer. I’m a business operator with 15+ years across product, marketing, analytics, and technical problem-solving, learning how to use AI well — with clear judgement, honest testing, and useful outcomes.

Personal

A practical AI learner with a business operator’s sense of what is useful.

I’m a Sydney-based Microsoft data and insights marketer, and a practical AI learner.

My background is in turning messy commercial questions into clearer decisions: dashboards, models, forecasts, customer insights, go-to-market programs, and operating rhythms. The common thread has always been translation — from business problem to useful system.

Lately, that system-building has expanded into AI-assisted tools. Some are for household life, some for health and reflection, some for content, and some for learning how modern AI workflows actually behave outside a demo.

The site is not a CV. LinkedIn can do that. This is a working notebook of the problems I’m choosing, the systems I’m shaping, and the lessons I’m learning as AI becomes part of everyday work.

Microsoft data and insights marketer 15+ years product, marketing, analytics and operations Azure Data Fundamentals MBA AGSM MITx Statistics & Data Science AI-assisted builder

System

My personal AI operating system.

Behind the projects is a small personal operating system I call PKA. It is not a product I am selling. It is the way I organise work with AI: specialist agents, clear source-of-truth files, repeatable workflows, automation, and public-safe proof-of-work.

I own

  • problem selection and workflow direction;
  • product judgement, privacy boundaries and usefulness criteria;
  • testing in the real target medium;
  • deciding what is good enough to keep, change, or throw away.

AI helps with

  • implementation drafts, code and configuration options;
  • documentation and debugging;
  • alternative approaches;
  • critique and edge-case checks.

Projects

Small systems, real use, public-safe case notes.

Each project starts from a practical problem. Published work uses sanitized text, demo data, simplified diagrams and public-safe repositories rather than private source material.

Published showcase

YouTube Digest

Problem

Too many new videos, limited attention, and useful ideas buried inside long transcripts.

Role

Misha defined the household use case, guided discovery and transcript handling, shaped the summary and email output, and set the privacy boundary.

Result

A weekly AI-assisted digest that helps a household choose which long-form YouTube episodes are worth watching.

PublishedGitHubContent workflow